Guitard Dominic, Saint-Aubin Jean, Tehan Gerald, Tolan Anne
École de psychologie, Université de Moncton, Moncton, New Brunswick, E1A 3E9, Canada.
University of Southern Queensland, Toowoomba, Queensland, Australia.
Mem Cognit. 2018 Feb;46(2):244-260. doi: 10.3758/s13421-017-0761-9.
In short-term serial recall, it is well-known that short words are remembered better than long words. This word length effect has been the cornerstone of the working memory model and a benchmark effect that all models of immediate memory should account for. Currently, there is no consensus as to what determines the word length effect. Jalbert and colleagues (Jalbert, Neath, Bireta, & Surprenant, 2011a; Jalbert, Neath, & Surprenant, 2011b) suggested that neighborhood size is one causal factor. In six experiments we systematically examined their suggestion. In Experiment 1, with an immediate serial recall task, multiple word lengths, and a large pool of words controlled for neighborhood size, the typical word length effect was present. In Experiments 2 and 3, with an order reconstruction task and words with either many or few neighbors, we observed the typical word length effect. In Experiment 4 we tested the hypothesis that the previous abolition of the word length effect when neighborhood size was controlled was due to a confounded factor: frequency of orthographic structure. As predicted, we reversed the word length effect when using short words with less frequent orthographic structures than the long words, as was done in both of Jalbert et al.'s studies. In Experiments 5 and 6, we again observed the typical word length effect, even if we controlled for neighborhood size and frequency of orthographic structure. Overall, the results were not consistent with the predictions of Jalbert et al. and clearly showed a large and reliable word length effect after controlling for neighborhood size.
在短期系列回忆中，众所周知，短单词比长单词更容易被记住。这种词长效应一直是工作记忆模型的基石，也是所有即时记忆模型都应考虑的基准效应。目前，对于是什么决定了词长效应尚无共识。贾尔伯特及其同事（贾尔伯特、尼斯、比雷塔和苏普雷南特，2011年a；贾尔伯特、尼斯和苏普雷南特，2011年b）认为邻域大小是一个因果因素。在六个实验中，我们系统地检验了他们的观点。在实验1中，使用即时系列回忆任务、多种词长以及一大组控制了邻域大小的单词，出现了典型的词长效应。在实验2和3中，使用顺序重建任务以及具有多个或少量邻域的单词，我们观察到了典型的词长效应。在实验4中，我们检验了这样一个假设，即之前在控制邻域大小后词长效应的消失是由于一个混淆因素：正字法结构的频率。正如所预测的，我们使用了正字法结构频率低于长单词的短单词，从而逆转了词长效应，就像贾尔伯特等人的两项研究中所做的那样。在实验5和6中，即使我们控制了邻域大小和正字法结构的频率，我们再次观察到了典型的词长效应。总体而言，结果与贾尔伯特等人的预测不一致，并且在控制邻域大小后清楚地显示出了巨大且可靠的词长效应。
